Can a NJ Solo Attorney Ethically Use AI to Prepare Client Wills and Estate Plans?

Solo attorneys who handle estate planning in New Jersey are increasingly getting pitched by document automation tools that promise to cut drafting time in half. Some of these tools are purpose-built for wills, powers of attorney, and healthcare directives. Others are general-purpose AI platforms being adapted for the job. The pitch is simple: feed in client intake data, get a polished draft, review and sign off.

The pitch is also incomplete.

What no vendor brochure spells out is where New Jersey's Rules of Professional Conduct intersect with that workflow, and what "review and sign off" has to mean in practice to stay on the right side of RPC 1.1.

What RPC 1.1 Actually Requires When AI Does the Drafting

RPC 1.1 requires competence, which the rule defines as the legal knowledge, skill, thoroughness, and preparation reasonably necessary for the representation. For estate planning, that's always included knowing the document. Knowing not just that the clauses are present, but that they work together, reflect current New Jersey law, and match what the client actually needs.

When an AI tool generates a will or a durable power of attorney, the competence obligation doesn't transfer to the software. It stays with the attorney. That sounds obvious, but the practical failure point isn't attorneys who skip review entirely. It's attorneys who review at too high a level, catching surface errors while missing structural ones.

A few categories of error that AI document tools commonly produce in estate drafts:

• Residuary clause mismatches. A tool trained on multi-jurisdictional templates sometimes generates a residuary clause that conflicts with specific bequests named earlier in the same document. It reads fine on a quick pass. It creates a dispute after death.
• Outdated NJ Uniform Trust Code provisions. New Jersey's trust law has been amended, and AI tools trained on older corpora can produce trustee powers language that no longer reflects current NJUTCA defaults. The clause isn't wrong on its face; it's just redundant or misleading in a way that matters.
• Beneficiary designation blind spots. AI drafting tools work on the document in front of them. They don't ask about the client's IRA, 401(k), or life insurance policy. A solo attorney using one of these tools without running through a full asset inventory may deliver a will that's internally clean but functionally incoherent with the client's broader estate.

None of these errors require an AI to malfunction. They happen when the attorney treats the AI output as a competent first draft rather than a starting point that still needs full legal analysis applied to it.

The Intake Data Problem

Most AI estate planning tools work off structured intake forms. The quality of the output depends almost entirely on the quality of the data going in. For a straightforward estate, that's a manageable risk. For anything with blended families, prior marriages, business interests, minor beneficiaries, or special needs considerations, the intake form is simply not going to capture enough nuance to drive a reliable draft.

The practical problem for NJ solo attorneys is that the "simple" estate is rarer than it looks at intake. A client who presents as a single person with no kids and a modest estate may have a disabled sibling they want to protect without disqualifying them from Medicaid. A married couple with adult children from prior marriages may not volunteer information about unequal distributions unless specifically asked.

AI tools do not probe. They populate. That gap is the attorney's job to close before anything gets drafted, not after.

What a Competent Workflow Actually Looks Like

Using AI for estate planning document drafting isn't the problem. Skipping the steps that make the AI useful is.

A defensible workflow for a NJ solo attorney looks something like this:

First, conduct a substantive intake conversation (not just a form) that covers the full asset picture, family structure, and client priorities. The AI does not do this for you.

Second, use the AI tool to generate an initial draft from your intake data. Treat this draft the way you'd treat a first-year associate's work: assume it needs correction, not just approval.

Third, review the draft against current New Jersey statutes. For wills, that means Title 3B of the New Jersey Statutes. For powers of attorney, the New Jersey Uniform Power of Attorney Act (N.J.S.A. 46:2B-8.1 et seq.). These are not optional checks.

Fourth, verify that the estate planning documents are consistent with any existing beneficiary designations, joint account arrangements, and trust structures. This requires information the tool doesn't have.

Fifth, document your review. If a disciplinary matter ever arises, the question will be whether you exercised independent professional judgment. A record of your review process is your answer to that question.

The Competence Standard Is Moving

The NJ Supreme Court and the ABA have both signaled that technological competence under RPC 1.1 includes understanding the tools an attorney uses. For New Jersey attorneys, that means knowing what your AI drafting tool was trained on, how it handles New Jersey-specific law, and where it has documented limitations.

Asking a vendor those questions before you commit to a tool isn't being overly cautious. It's the minimum due diligence the competence rule now expects. Most vendors won't answer with precision. That answer itself tells you something important about how much weight to put on their output.

If you're evaluating an AI estate planning tool right now, request the vendor's documentation on jurisdiction-specific training data and any published accuracy benchmarks on trust and estate documents. If they can't produce either, build your review workflow with that uncertainty priced in.